Rain Dances in the Dry Season:
Overcoming the Religious Congruence Fallacy
Mark Chaves
Department of Sociology Duke University [email protected]
Published in the Journal for the Scientific Study of Religion 49(1):1-14, March, 2010. The definitive version is available at www.blackwell-synergy.com.
This article was the basis of my Presidential Address, delivered on October 24 at the SSSR’s 2009 annual meetings in Denver. Many thanks to those who offered helpful comments on earlier versions: Gary Adler, Kevin Christiano, Jay Demerath, Art Farnsley, Linda George, Kieran Healy, Janet Jacobs, Kay Meyer, Andrew Miles, Margarita Mooney, Phil Morgan, Mark Regnerus, Chris Smith, Lynn Smith-Lovin, Ed Tiryakian, Grant Wacker, Steve Warner, and Rhys Williams. Earlier versions were helpfully discussed by the Community of Scholars at the Center for Spirituality, Theology, and Health, the American Religious History Colloquium, and the Department of Sociology, all at Duke University. Special thanks to Art Farnsley for suggesting the phrase, “religious congruence fallacy.”
Rain Dances in the Dry Season
Overcoming the Religious Congruence Fallacy
ABSTRACT
Religious congruence refers to consistency among an individual’s religious beliefs and attitudes, consistency between religious ideas and behavior, and religious ideas, identities, or schemas that are chronically salient and accessible to individuals across contexts and situations. Decades of anthropological, sociological, and psychological research establish that religious congruence is rare, but much thinking about religion presumes that it is common. The religious congruence fallacy occurs when interpretations or explanations unjustifiably presume religious congruence.
I illustrate the ubiquity of religious incongruence, show how the religious congruence fallacy distorts thinking about religion, and outline an approach to help overcome the fallacy.
2
RAIN DANCES IN THE DRY SEASON
Overcoming the Religious Congruence Fallacy
After reading a book or article in the scientific study of religion, I wonder if you ever find yourself thinking, “I just don’t believe it.” I have this experience uncomfortably often, and I think it’s because of a pervasive problem in the scientific study of religion. I want to describe that problem and how to overcome it.
The problem is illustrated in a story told by Meyer Fortes. He once asked a rainmaker in a native culture he was studying to perform the rainmaking ceremony for him. The rainmaker refused, replying, “Don’t be a fool, whoever makes a rain-making ceremony in the dry season?”
(Tambiah 1990:54).
The problem is illustrated in a different way in a story told by Jay Demerath. He was in
Israel, visiting friends for a Sabbath dinner. The man of the house, a conservative rabbi, stopped in the middle of chanting the prayers to say cheerfully: “You know, we don’t believe in any of this. But then in Judaism, it doesn’t matter what you believe. What’s important is what you do”
(Demerath 2001:100).
And the problem is illustrated in yet another way by the Divinity School student who told me not long ago that she was having second thoughts about becoming an ordained minister in the
United Church of Christ because she didn’t believe in God. She also mentioned that, when she confided this to several UCC ministers, they told her not to worry about it since not believing in
God wouldn’t make her unusual among UCC clergy.
These stories illustrate in different ways a problem long recognized by social psychologists and cultural analysts: attitudes and behavior correlate only weakly, and collections
3 of apparently related ideas and practices rarely cohere into logically unified, mutually reinforcing, seamless webs (Maio et al. 2003, Swidler 1986, DiMaggio 1997, Vaisey 2009).
Instead, ideas and practices exist as bits and pieces that come and go as situations change, producing many inconsistencies and discrepancies. This is true of culture in general, and it is true of religious culture in particular. Observant Jews may not believe what they say in their
Sabbath prayers. Christian ministers may not believe in God. And people who regularly dance for rain don’t do it in the dry season.
I will use “religious congruence” in three related senses: (1) individuals’ religious ideas constitute a tight, logically connected, integrated network of internally consistent beliefs and values; (2) religious and other practices and actions follow directly from those beliefs and values; and (3) the religious beliefs and values individuals express in certain, mainly religious, contexts are consistently held and chronically accessible across contexts, situations, and life domains. In short, it can mean that religious ideas hang together, that religious beliefs and actions hang together, or that religious beliefs and values indicate stable and chronically accessible dispositions in people.
These are not identical meanings, but I combine them in this overarching concept of congruence because the evidence overwhelmingly shows that people’s religious ideas, values, and practices generally are not congruent in any of these senses. Rather, people’s religious ideas and practices generally are fragmented, compartmentalized, loosely connected, unexamined, and context-dependent. This is not a controversial claim; it’s established knowledge. But this established knowledge does not inform our research and thinking as centrally and deeply as it should. Even though we know that cultural fragmentation is common, we often interpret our research findings in ways that presuppose a congruence that we know is not generally there.
4
This is the religious congruence fallacy, and it pervades the scientific study of religion. This is why I often think, “I just don’t believe it,” when I read work in the scientific study of religion.
Religious incongruence is not the same thing as religious insincerity or hypocrisy. I am not saying that the rain dancer or the rabbi or the UCC clergy are religious hypocrites. On the contrary, they are the heroes of this story because they illustrate something true about religion in general. They don’t commit the religious congruence fallacy. We commit the religious congruence fallacy when we fail to heed the lesson they teach us.
I do three things in this article. First, I illustrate the ubiquity of religious incongruence.
Second, I show how conventional thinking in the scientific study of religion and beyond presumes just the opposite—that religious congruence is common rather than rare. Third, I describe an approach aimed at reducing the religious congruence fallacy’s influence on the scientific study of religion.
The Ubiquity of Religious Incongruence
Ideas and practices mainly are fragmented and situational rather than congruent. This is well-established knowledge in anthropology, sociology, political science, and psychology
(Converse 1964, Evans-Pritchard 1965, Swidler 1986, D’Andrade 1995, DiMaggio 1997, Maio et al. 2003, Vaisey 2009). I don’t think anyone ever has marshaled evidence to make this point about religion in particular, so I want to begin by describing examples and evidence illustrating that religion is no exception to the generalization that people’s ideas and actions do not usually cohere into tightly connected wholes.
One way to see religious incongruence clearly is to examine the most instrumental- looking ritual and religious action. The key observation is that instrumental-looking ritual and
5 religious action usually supplements practical action, even when congruence would lead us to expect it to replace practical action. Ludwig Wittgenstein articulated this point when he commented on James Frazer’s The Golden Bough, which exudes congruence by assuming, for example, that people believe that stabbing an enemy’s effigy before battle guarantees military success, or that morning rituals make the sun rise. In considering these examples, Wittgenstein pointed out that the same person who stabs an effigy also carefully crafts and sharpens his weapons. About dawn rituals he said, “The rites of dawn are celebrated by the people, but not in the night; rather there they simply burn lamps” (Tambiah 1990:56-58).
E. E. Evans-Pritchard made a similar point with this example: “[S]ome peoples put stones in the forks of trees to delay the setting of the sun; but the stone so used is casually picked up, and has only a mystical significance in, and for the purpose and duration of, the rite. The sight of this or any other stone in any other situation does not evoke the idea of the setting sun.
The association . . . is brought about by the rite, and need not in other situations arise” (1965:88-
89). His more general point was this: “All observers who have made lengthy first-hand studies of primitive peoples are agreed that [these people] are for the most part interested in practical affairs, which they conduct in an empirical manner, either without the least reference to supra- sensible forces, influences, and actions, or in a way in which these have a subordinate and auxiliary role” (88). Religious and practical beliefs “in reality are found in different situations and at different levels of experience [and] mystical representations are not necessarily aroused by objects outside their use in ritual situations” (88-89). Instrumental-looking rituals and religious action and belief usually are performed and expressed only in specific contexts, and religious action almost always accompanies rather than replaces pragmatic belief and action.
6
Far from being a feature only of primitive religion, this is true wherever there is instrumental-looking ritual and religious action. Many people believe in divine healing and actively seek it, for example, and prayers and anointing for healing, and testimony about divine healing, are common at Pentecostal and other worship services. But very few people seek divine healing instead of medical treatment--unless, of course, they don’t have access to quality medical care (Poloma 1989:54-60; Wacker 2001:27, 191). Divine healing testimonies often contain the exclamation, “The doctors were amazed!”, again indicating that healing prayers and rituals mainly supplement rather than replace medical care, just as superstitious athletes can believe that rituals or talismans improve their play while they also train and practice incessantly. People all the time pray for health or wealth or victory in battle, and such instrumental-looking religious action may look like dancing for rain in the dry season. But the fact that such action almost always supplements rather than replaces practical action shows otherwise.
Incongruence also is evident in ritual and religious action that is not instrumental. Here, the telling fact is that the realities expressed in rituals, the beliefs and attitudes that people express outside of ritual, and the practical realities of peoples’ everyday lives usually only loosely connect with one another.
One of my favorite ritual analyses is Barbara Myerhoff’s (1978) account of a graduation ceremony at a Jewish senior citizens center in Venice, California. The graduation ceremony marked the completion by 26 elderly Jews of a five-month course in Yiddish history. In this instance, the relationship between the ritual content and the participants’ everyday lives was closer to outright contradiction than to congruence. Most poignantly, the ceremony celebrated the love and honor bestowed on these senior citizens by their children and grandchildren. The ceremony portrayed the graduates as having “realized their most cherished ideals in life by
7 producing children who were educated, successful, and devoted to them” (106). One speaker said this to the gathered crowd: “How proud you must be of your parents and grandparents this day. How you can rejoice over them. . . You, the children they have produced, are their greatest accomplishment, and you are here to show them you understand what they have done” (94). But the audience contained almost no children or grandchildren of the graduates. And a major, excruciating, part of these people’s everyday lives was their feeling of being neglected and unappreciated by their highly successful but rarely visiting progeny.
In this way, and in other ways, this ritual was “built upon contradictions, denials, and fictions” (86). It disguised rather than reflected everyday realities. None of this is to say that this was an unsuccessful or inauthentic ritual, or that it produced no solidarity or positive emotional energy. By any reasonable standard it was a highly successful event, and the contradictions do not diminish that success. How rituals work their magic and achieve success is a subject for another day. Here, the point is that it is not because they connect tightly, coherently, and literally to the beliefs, attitudes, and practices evident in people’s lives outside the ritual.
Myerhoff analyzed a one-time, unique ritual rather than a regularly occurring event that has been polished over the years, but religious incongruence clearly is not limited to one-time events. Timothy Nelson was struck that a poor black congregation held long, emotionally intense worship services emphasizing love of one’s fellow Christians while, outside of worship, members often expressed distrust, harsh criticism, and negative feelings about each other
(Nelson 2005:135, 182-86). Cheryl Townsend Gilkes described a church service characterized by “extreme emotionalism” in which “men and women scream and cry and leap about. Bodies seem to be wracked by uncontrollable spasms of both grief and joy. People, both young and old,
8 leap about in the aisles, dance at the altar, and fall out on the floor . . . Choir members who ‘fall out’ during the performance are supported by other choir members [and choirs sing] six or seven choruses of a song until the lead singer regains consciousness in order that she may finish the song with the choir properly” (1980:38-39). Gilkes goes on: “No matter how severe the pandemonium within the church service, I have never witnessed a church service in which every single person’s episode of ‘getting happy’ or ‘shouting’ was not resolved, worked through, or finished before the singing of the final hymn and the recessional. When the participants leave, they usually appear as unruffled as they did when they came into the church” (39). Grant
Wacker made the same point about white Pentecostals, who, he says, “almost certainly fell into physiologically disassociative states but . . . they also knew exactly how and when to enter and leave those states” (Wacker 2001:56). On the other end of the social and religious spectrum, only two-thirds of church-going Episcopalians say that they know that God exists, but the other third still say, “Our Father, who art in heaven,” at the appropriate moment in the liturgy.
Religious action—instrumental or not—is highly situational.
Religious behavior is not the only type of behavior that is strongly situation dependent.
Steven Hitlin said this in his 2008 book on the social psychology of conscience: “Decades of social psychological research can be boiled down to one insight: if we want to predict someone’s behavior, we are better off knowing where they are rather than who they are” (Hitlin 2008:93).
Erving Goffman would not have been surprised by this, and the power of the situation applies in spades to religion. Religious ideas generally are activated in people’s minds only situationally, and religious action, including the act of expressing religious beliefs or attitudes, mainly is bounded and situational. To quote Evans-Pritchard again, the associations that people make among religious ideas and actions occur “only when evoked in specific ritual situations, which
9 are also of limited duration” (1965:29). David Smilde recently drew the same conclusion from his study of Venezuelan evangelicals: “[I]n most contexts the religious laity have a repertoire of meanings that are marshaled for specific situations. This repertoire will not normally be rationalized but rather will contain numerous mutually incompatible meanings that are inconsistently used” (Smilde 2007:125). And the authors of a 2008 Science article reached the same conclusion after reviewing research on the relationship between religiosity and prosocial behavior: When it comes to predicting prosocial behavior, “the religious situation is more important than the religious disposition” (Norenzayan and Shariff 2008:62, my emphasis).
Examples of religious incongruence could be multiplied endlessly. We see it across religions, cultures, historical periods, and social contexts. Franz Boas described a successful
Indian shaman who did not believe in his own powers (Palmer and Steadman 2004:140). The
Maasai people describe God concretely and anthropomorphically during their rituals, but they speak much more abstractly when asked in conversation what they believe about God (Olsson
1999), which makes them a lot like American churchgoers. Maine lobster fishermen “observed taboos that they denied believing in [and] failed to observe taboos in which they professed belief” (Palmer and Steadman 2004:140). Mary Jo Neitz interviewed a Wiccan priest who told her that it doesn’t matter whether he believes that his rituals work; what matters is doing them
(Neitz 2004:400). People who claim to be born-again or evangelical Christians are no less likely than others to endorse certain ideas they ought to be hostile to, such as reincarnation, channeling, or astrology (Gallup and Lindsay 1999:40). Among respondents to the General Social Survey, conservative Protestants are no less likely than other Protestants to have been divorced, to have seen an X-rated movie in the last year, or to be sexually active even if they aren’t married. And the reality of decision-making and economic responsibility in evangelical families is very far
10 from the ideals these same people espouse about the husband being the head of the household
(Gallagher and Smith 1999).
But my primary goal is not to persuade you that religious incongruence is common. As I mentioned earlier, the ubiquity of cultural incongruence in general, and religious incongruence in particular, is an established, uncontroversial fact. My primary goal is to persuade you that this established knowledge should more centrally and deeply inform our work. Despite knowing that religious congruence is rare, we often forget this in practice.
Presuming Congruence Rather than Incongruence
How does the religious congruence fallacy shape the scientific study of religion? Most obviously, the religious congruence fallacy has inspired a search for religious influence on behavior that religious congruence implies should be there, and the most common form in which the religious congruence fallacy appears is when we explain behavior by connecting it to religious affiliations, practices, or beliefs that seem consistent with it and from which the behavior is thought to derive. But decades of research devoted to proving that religiosity is consequential in ways that congruence implies has produced a confusing hodge-podge of mixed results. This should not be surprising. Mixed results are exactly what we should expect if religious incongruence is ubiquitous.
Mixed results appear whether we look at religiosity’s connections with health, political behaviors, sexual behavior, prosocial behaviors, antisocial behaviors, or other outcomes.
Religiosity, for example, seems related to positive health outcomes in the United States but not in Europe (King 2009). Another example: Evangelical Protestant pregnant teenagers are less likely than mainline Protestants to abort (controlling for other things), but Catholics are not—and
11 girls in religious schools are more likely to abort than girls in public schools (Adamczyk 2009).
After reviewing the literature on the connections between religion and prejudice and religion and deviance, Robin Perrin concluded that “religiosity effects are not as significant as common wisdom and social scientific theory might suggest” (Perrin 2000:537). And decades of psychological research looking for behavioral consequences of intrinsic religiosity has yielded the conclusion that intrinsically religious people do not act in more pro-social ways than anyone else, but they think they do, or should, or would, so their behavioral self-reports often are different from those of other people even when their behavior is not (Trimble 1997; Batson and
Powell 2003; Leach et al. 2008).
I do not have space to describe as many examples of mixed results as I would like, but I want to highlight one particularly compelling example. A very common finding in the scientific study of religion is that theological beliefs relate to other beliefs and actions in different, sometimes opposite, ways for African Americans than for whites. Controlling for other things, theologically conservative whites are more politically conservative than theologically liberal whites, but the opposite is true for African Americans (Greeley and Hout 2006: Chapter 4).
White conservative Protestants are more individualistic when they think about inequality, but the opposite is true for African Americans (Emerson and Smith 2000:97). Theologically conservative white congregations are less socially engaged than theologically liberal white congregations, but the opposite is true among African American congregations (Tsitsos 2007).
Among white adolescents, religiosity (though not evangelicalism) is associated with delayed sexual activity, but “this is much less true for African Americans” (Regnerus 2007:152). White conservative Protestants give more to their churches than other Protestants and Catholics do, but black conservative Protestants do not give more to their churches even though their religious
12 beliefs and attitudes are essentially the same as white conservative Protestants (Keister
2008:1260-61).
This very common finding—religious beliefs and attitudes relate differently to other beliefs and behaviors for whites than they do for blacks—is decisive evidence of religious incongruence, and it alone should make us reluctant to explain religious and other behavior by connecting that behavior to religious affiliations, practices, and beliefs from which the behavior is thought to logically derive. If congruence between religious beliefs and other beliefs and behaviors in fact produced the correlations we observe among whites, the same correlations should appear among African Americans when other relevant variables are controlled. But they do not. We know this, but we have not appreciated its theoretical significance: religious incongruence is ubiquitous, and it usually will be a mistake to connect religiosity to what look like logically related outcomes. Almost every claim of the form, “People act in a certain way because they are in a particular religion or because they attend religious services or because they hold this or that religious belief,” commits the religious congruence fallacy.
It always is possible to develop interpretations that account for a complex pattern of mixed results. But the complicated interpretations that are required to account for mixed results in the study of religion remind me of the epicycles that Ptolemaic astronomers posited to make the facts about how planets moved in the sky consistent with an earth-centered cosmology.
Positing that planets moved in complex loops—epicycles—made the observations consistent with an earth-centered cosmology. But these epicycles’ complexity now seems a clear sign that something was fundamentally wrong with the ideas behind them. Similarly, research results filled with apparent contradictions that require interpretations loaded with caveats and exceptions and nuances should lead us to consider the possibility that something is wrong with our basic
13 ideas. I think what is wrong is the presumption of religious congruence. The religious congruence fallacy leads us to search for causal effects we should not be searching for, it leads us to make claims about religious causality we should not make, and it leads us to posit epicycles to save the religious congruence assumption from evidence that contradicts it.
The quantitative search for “religious effects” is particularly vulnerable to the religious congruence fallacy, but this fallacy is not reducible to unjustified causal inference in quantitative work. It occurs commonly in both quantitative and qualitative research. Its telltale sign in quantitative research is a regression model in which the coefficients attached to religious service attendance, religious belief, or religious affiliation are interpreted causally. Its telltale sign in qualitative work is when beliefs or attitudes that a researcher hears in the field are treated as stable dispositions presumed to be equally salient across situations, when people’s accounts of how their actions are rooted in their beliefs and values are treated as causal explanations of those actions, or when the coherent stories we all try to tell about ourselves are presumed to reflect real congruence in our everyday lives. Congruence is a suspect explanation of behavior whether it is imputed by an analyst or produced by an interviewee. The religious congruence fallacy is conceptual, not methodological, so the difference between a survey and an in-depth interview is not fundamental; nor is the difference between a regression equation and a qualitative analysis.
In either case, if expressions of religiosity are taken to indicate stable, pan-situational, dispositions with logically clear causal connections to other beliefs or to actions, then the religious congruence fallacy lurks.
The interpretive problem I am emphasizing is not unique to religion. Similar issues arise when thinking about the meaning of gender and race effects. But the congruence fallacy is particularly problematic in the scientific study of religion because interpretations that rest on
14 congruent beliefs and ideologies more readily come to mind when we think about religion than when we think about gender or race or other independent variables, so it is particularly easy for scholars of religion to fall back on congruence as a default interpretation, even when they shouldn’t.
I want to be clear about something I am not saying. I am not saying that religious congruence is impossible. I am saying that it is rare, and much conventional practice does not appreciate how rare it is. Religious congruence is rare because achieving it in a specific situation requires one or more of three conditions.
Congruence can be achieved through conscious cognitive effort. We can analyze our different ideas’ relations to each other and to action, and we can try to reduce recognized inconsistencies. But this does not come naturally, and it is hard to do. That is why it is rare.
Congruence also can be achieved through social rather than cognitive effort. I can consult a religious leader about what my religion demands in a specific situation, and I can act according to his advice, thereby achieving congruence through deference to an authority.
Congruence also can be achieved socially through immersion in a homogeneous religious culture or through intense involvement with an all-encompassing group. But this sort of deference to religious authority or enveloping religious culture is unusual. Mark Regnerus studied religious influence on teenage sexuality, but I think he makes a point that is more generally true when he says that religious influence occurs only when people are embedded in a “a network of like- minded friends, family, and authorities,” and that such embeddedness is “relatively rare,
[encompassing only] a small segment of American adolescents, excluding even most religious adolescents” (Regnerus 2007:203-4). Sacred canopies exist, but they are rare.
15
A third, and probably more important, way that religious congruence can be achieved is through experience that forges internalized, automatic responses to situations so that religious schemas spring automatically to mind in certain situations. When internalized responses or schemas are firmly in place, religious congruence can occur without cognitive effort or social control. Internalization promotes what social psychologists call heuristic processing and what cognitive psychologists call connectionist or parallel processing. Whatever the labels, the basic idea is the same: some human information processing is “implicit, unverbalized, rapid, and automatic, [and some is] explicit, verbalized, slow, and deliberate” (D’Andrade 1995:180).
Internalization promotes the former.
Internalization is the most powerful route to congruence, but it too is difficult to achieve.
Roy D’Andrade (1995:144) notes that there is a trade-off between how easy it is to learn a schema and how deeply that schema is internalized. Serial or systematic learning--the proposition-based, verbal kind of learning--is relatively easy. But the knowledge gained through serial learning is difficult to internalize in ways that allow it to shape our behavior without having to go through the laborious process of calling content to mind and consciously deciding what it implies about action in a particular situation. Learning that leads to heuristic or connectionist processing, on the other hand, usually is much more time-consuming. It can require hundreds or even thousands of experiential repetitions--not, importantly, mere verbal repetitions--to make a response to a situation automatic and largely unconscious.
To use D’Andrade’s example, an American traveling in England learns very quickly that they drive on the left over there. This is explicit, conscious learning, and it is relatively easy.
But it takes much longer to do the correct thing automatically when you come to a roundabout or
16 cross a street. Habitual knowledge about driving on the left comes only from many driving experiences. Reading the guidebook 1,000 times won’t help.
Repeated experience is one route to internalization, but internalized, automatic responses do not always require hundreds of repetitions to create. Anyone who feels queasy at the smell of food with which they once had a bad experience knows that. But intense experiences that can create this sort of automatic stimulus-response connection are difficult to create and control.
Whether internalization occurs through many repeated experiences or through a single intense experience, it is a kind of learning that requires special circumstances.
The religious congruence fallacy often amounts to assuming that religious beliefs and attitudes combine the easy learnability of self-consciously acquired knowledge with the easy causal influence of internalized, automatic knowledge. We imagine that, because religious people hear hundreds of sermons connecting religious faithfulness with, say, caring for the poor, or hear many exemplary religious stories about caring for the poor, or participate in many rituals in which caring for the poor is emphasized, religious people therefore will be more likely to help the poor when given the opportunity to do so outside the religious setting. This model of religious influence assumes that systematic learning--the explicit, verbal, kind of learning-- produces internalized connections that we then act on more or less automatically. But religious schemas do not become chronically accessible and influential because they are memorable, commonly enacted in rituals, or repeated often in religious narratives. Internalized religious schemas that produce automatic responses in other settings are unlikely to form without repeated experiences in religious settings that are reproduced more or less exactly in other settings.
People generally need many experiences akin to really driving on the left to forge the relevant
17 internalized responses. Neither repeated exposure to relevant narratives nor repeated participation in rituals is enough to generate automatic processing.
It is important to emphasize religion’s situational specificity again here. Repeated religious practice clearly establishes many deeply internalized responses in people. But because religious practice mainly is situation specific, the automatic responses it produces also mainly will be situation specific. A lifetime of Easter services or Passover Seders or Ramadan fasting or rosary praying surely establishes powerful, persuasive, long-lasting moods and motivations specifically regarding Easter or Passover or Ramadan or the rosary, but these internalized connections do not necessarily extend beyond Easter or Passover or Ramadan or the rosary. A lifetime of weekly churchgoing surely establishes internalized responses to certain hymns or stories or rituals or practices, but these internalized responses do not necessarily extend beyond the religious setting. A single emotionally intense experience at a youth camp may well produce a lifetime association by which someone is brought to tears whenever they hear a particular song, but that internalized response does not necessarily extend beyond that very specific association.
This is why the causal influence of religious ideas or practices rarely extends beyond the religious domain. Even when people have internalized a religious schema, it exists in our brains along with many other schema. To influence behavior in a given situation, it is not enough for a religious schema to be internalized. It also has to become activated in that situation.
So religious congruence obviously exists, but it requires substantial cognitive effort, intense and consistent social reinforcement, or internalization. The systematic learning we do in religious contexts does not generally produce automatic religious responses, and the internalizations that religious people forge through religious practice generally stay confined to religious contexts. To paraphrase a recent slogan enticing people to visit Las Vegas, what
18 happens in church mainly stays in church. This is why we commit the religious congruence fallacy almost every time we explain behavior by connecting it to religious affiliations, practices, or beliefs that seem consistent with it.
Another way that a presumption of religious congruence shapes the literature occurs when we consider congruence, or the drive towards congruence, to be religion’s defining feature.
Clifford Geertz defined religion like this, but it is a misleading way to think about religion. It encourages us to overlook the fact that, in everyday life, the powerful, persuasive, long-lasting moods and motivations associated with religion usually are specific and situation-dependent rather than general. And even the impulse to strive for congruence varies considerably across and within religions. An impulse towards this-worldly ethical rationalization is central to some kinds of Christianity, a point that Max Weber famously emphasized, but it is not central to all religions, and it is not uniformly central within Christianity. To mention just one classic example, surely not atypical, of Christianity without this impulse, Liston Pope observed in his
1942 classic, Millhands and Preachers, that, to clergy and members of “uptown churches,” “the private life of church members was none of the minister’s business . . . For uptown people, religion — well, it’s just religion — which is to say, it is a set of actionways and thoughtways associated with, and largely confined to, the church, with the minister as exemplar and chief practitioner” (93). Some religious leaders and people strive for religious congruence, but not all do. Striving for congruence is not an essential feature of religion—unless we declare it such by definition.
The religious congruence fallacy is widespread in the scientific study of religion, but it also reverberates more widely. Indeed, attributing an unwarranted coherence to people’s religious ideas and practices may be the single most important misunderstanding about religion
19 among the general public. It is behind both unrealistic popular idealizations of religion and unjustified popular demonizations of religion.
The Bush Administration’s faith-based initiative is a recent example of an unrealistic popular idealization of religion. I was once at a meeting at which Catholic Charities executives, who also were nuns, were told by faith-based initiative advocates that a new day had dawned and they no longer needed to keep religious content out of their programs. They now had the green light to infuse their programs with religion. But the executives were puzzled by this. They asked, “Why would we want to infuse our programs with religion?” The faith-based initiative was based in part on the conviction that it is sensible to do rain dances in the dry season, but it also was based on the conviction that religious people generally believe that it’s sensible to dance for rain in the dry season, and so they would welcome the invitation to do so. Most leaders of religious social services agencies know better, which is one reason the faith-based initiative went basically nowhere.
The religious congruence fallacy also underlies unjustified demonizations of particular religions. Western commentators commonly portray Islamic societies and Moslem people as societies and people in which religious and political beliefs are fused into a coherent whole that produces authoritarianism and extremism. But surveys of the Islamic world show that, despite considerable uniformity within and among Islamic countries in basic religious beliefs, there is considerable variation on many important social and political issues (Moaddel 2007:7).
Moreover, like people everywhere, citizens of Islamic countries are “simultaneously and sincerely attached to values that clash,” such as agreeing that “a good government is the one that makes law according to people’s wishes, while at the same time [agreeing] that a good
20 government implements only the shari’a laws” (Moaddel 2007:11). If we understood better the ubiquity of religious incongruence we might be less likely to demonize other religions.
The religious congruence fallacy also infects “new atheist” critiques of religion.
Christopher Hitchens’ nearly 300-page attack on religion contains many irrefutable criticisms, but some of these criticisms are compelling only as criticisms of religion that strives for excessive congruence rather than as criticisms of religion in general. Like many others both inside and outside the academy, Hitchens incorrectly considers this drive a defining feature of religion: Religion, he says, “must seek to interfere with the lives of nonbelievers, or heretics, or adherents of other faiths. It may speak about the bliss of the next world, but it wants power in this one” (Hitchens 2007:17, original emphasis). And: “religion is ultimately incapable of
[leaving unbelievers alone]. As I write these words, and as you read them, people of faith are in their different ways planning your and my destruction, and the destruction of all the hard-won human attainments that I have touched upon” (13). This is wrong, and it is wrong because it treats the drive for congruence as the defining feature of religion.
Daniel Dennett’s analysis of religion is more subtle, but it too relies on the religious congruence fallacy. Dennett (2006, chapter 8) correctly observes that most people do not act fully consistently with their religious beliefs. This is just another way to say that religious incongruence is common. But he draws the wrong conclusion from this observation. He says that people who do not act according to their professed religious beliefs, and people who talk about their religion differently in a ritual context than they do in casual conversation, act as though they do not really believe in God, even though they say they do. What such people really believe, he says, is that it is a good thing to profess belief in God even if you do not really believe in God. As he put it in an interview, “many people who profess belief in God do not
21 really act the way people who believe in God would act; they act the way people would act who believed in believing in God” (Floyd 2000). This is misleading. I would say that such people act the way competent social actors act: whatever they do or do not believe, they talk and act differently in different situations. Dennett’s response to religious inconsistency is to declare
“gotcha!” because he thinks religious incongruence is equivalent to religious confusion. But something is wrong with an analysis that leads to the conclusion that almost no one really believes in God. What’s wrong is forgetting that, like other action, the appropriateness of religious action, including the action of professing religious belief, depends on the situation.
Dennett surely knows this is true of action in general, but he commits the religious congruence fallacy by forgetting that it also is true of religious action.
Overcoming the Religious Congruence Fallacy
How might we overcome the religious congruence fallacy? Obviously, we should hesitate to treat religious beliefs as stable dispositions, we should hesitate to explain behavior by connecting it to religious affiliations, practices, or beliefs from which the behavior seems to follow, and we should try to better understand the conditions under which religious congruence really does emerge. But I want to go beyond these obvious implications.
I also want to go beyond the usual response to the problem of religious incongruence among cultural sociologists and anthropologists, which is to limit attention to practices, narratives, material objects, and other directly observable cultural objects that are external to people’s minds. As Paul Lichterman recently described this intellectual development, “Rather than gauging private beliefs, recent works advocate for investigating religious vocabularies and forms of self-presentation that we can see and hear in everyday life, and seeing what actions
22 accompany them” (Lichterman 2008:84). Mary Jo Neitz characterized this turn as one that
“moves us away from thinking of religion primarily in terms of beliefs and rules, and towards thinking in terms of practices” (Neitz 2004:399). The anthropologists Craig Palmer and Lyle
Steadman have gone even farther, calling for replacing wholesale the study of religious belief with the study of religious language (Palmer and Steadman 2004). We have learned much from research inspired by this turn away from mental states and towards directly observable cultural objects like discourse and practices. But we also lose something, especially in the study of religion. As Robert Wuthnow put it, “Religion is, after all, centrally concerned with beliefs and convictions [and] with the ways in which meaning and purpose are constructed . . . To avoid focus on these aspects of religion would be like trying to understand apple pie without paying attention to apples” (1997:253). Wuthnow aimed this barb at approaches that ignored religious culture in general, but it also applies to approaches to religious culture that avoid grappling with mental states.
This strikes close to home for me because ignoring belief and other mental states has been my own response to the fact of religious incongruence. This is why I conceptualized secularization as the declining scope of religious authority rather than declining religion. This is why I studied rules about women’s ordination rather than female religious leaders. This is why I was drawn to the National Congregations Study, a study of organizational characteristics and collective practices. Much like those who respond to the incongruence problem by turning away from the study of belief and conviction to the study of language and practices, I turned away from the study of belief to the study of religion’s social organization. I believed that religious incongruence implied that we could make little scientific progress by studying religious mental states.
23
But I have changed my mind about this, and I want to outline an approach that neither naively ignores the religious congruence problem nor turns completely away from the study of mental states. I seek middle ground in another sense as well: I have criticized conventional practice in ways similar to how constructionist or interpretive scholars might criticize it, but I do not respond to this problem by abandoning a truly scientific study of religion.
So, how might we overcome the religious congruence fallacy? First, anyone who offers an explanation that presumes religious congruence should bear a heavier burden of proof. In an ideal world, this would mean that such explanations would establish that the unusual conditions necessary for religious congruence exist in the situation at hand. If we want to claim religious internalization, we would have to show that people have had enough relevant, reinforcing religious experiences to forge an internalized connection strong enough to produce an automatic habitual response in a particular situation. We might call this the “would they refuse to eat it” standard of proof since several paradigm examples of truly internalized automatic religious responses involve food: Orthodox Jews and Muslims not eating pork, Seventh-day Adventists not eating meat, Hindus not eating beef, and so on. The religious congruence fallacy sometimes amounts to treating a connection between some sort of religion and some sort of behavior as if it were analogous to Orthodox Jews not eating pork. But this analogy is not apt for most of what we study. It seems unlikely, for example, that evangelical Protestants have automatic, internalized reactions to pornography or premarital sex, or that Catholics have automatic reactions to abortion, that are analogous to Orthodox Jews’ reactions to pork. In an ideal world, researchers whose explanations rely on internalization would establish that the connections they posit meet this standard and are activated in relevant decision-making situations. Similarly, in an ideal world, if someone thinks that religious congruence is achieved through conscious cognitive
24 processing or social control, then that person would be expected to establish that such conscious processing really is happening at key decision-making moments or that effective social control really does exist. If human action involves a mix of habit and creativity, then, in an ideal world, claims that religion influences action would be accompanied by evidence that religion really did shape the relevant habits or really does inform our creativity at decision-making moments.
That’s the ideal world. In the real world—where assistant professors need to publish papers even when relevant evidence is hard to come by, and where reviewers and editors require us to tell stories about correlations between religiosity and other things—shifting the burden of proof might mean that those who offer claims presuming congruence should be required to write a sentence something like this: “I know that religious congruence is very unusual, but here’s why my case should be considered an exception.” And the ensuing justification should say how congruence has been achieved in this case. This justification needs to persuade us that people really had many experiences akin to driving on the left, or really consciously reflect on religion at decision-making moments, or really live in a setting with effective religious social control.
Neither correlations nor official ideologies nor elite discourse nor ritual practice nor narrative content nor people’s post-hoc interpretations of their actions will help much in making this case.
And if it seems unlikely that enough people would have had enough relevant experiences to forge the necessary specific internalizations, or if it seems unlikely that the relevant internalizations, if they exist, would be activated in key situations, and if neither explicit intellectual effort nor effective social control seems likely, then we should recognize that religiosity is unlikely to be the causal force it may at first appear to be.
I do not want to abandon religion as an individual-level independent variable, but I do want us to less quickly resort to religious congruence mechanisms when we think about religious
25 causality. I am not saying that we should stop telling stories about our correlations or our field observations. We have to tell stories about our results. But I am saying that we should hesitate to tell stories that presume religious congruence, and we should expect those who tell such stories to justify them. Shifting the burden of proof in this way will help us overcome the religious congruence fallacy because the implausibility of most explanations that presume congruence will become apparent when the alleged congruence mechanisms are made explicit.
Second, we might take seriously mental states other than beliefs. The attempt to find connections between internalized religion and various outcomes has focused overwhelmingly on religious beliefs. But beliefs are just one kind of mental state. We also have perceptions, expectations, feelings, wishes, attitudes, and intentions, all of which might shape action.
Unconscious mental states also shape action. Social psychologists tell us that, when compared with attitudes and behavioral intentions, beliefs, especially general beliefs, are the least predictive of behavior (Lord 1997:235-249). So focusing on belief as the religious mental state par excellence makes us particularly vulnerable to the religious congruence fallacy. Examining religiously relevant mental states other than belief also will connect us more effectively to other literatures, such as the medical literature on placebo effects, which work mainly through expectations (Stewart-Williams and Podd 2004), or the vast psychological literature on priming, which shows that unconscious mental states shape behavior.
Third, we might pay more attention to decision-making situations. An internalized religious schema does not necessarily produce religious influence; that schema also has to be activated in a specific decision-making situation. We might investigate what features of situations activate religious schema in people. The psychological literature on priming could help here. Does religious jewelry or clothing prime religious ideas? How about seeing a
26 religious building on the corner, or being in the presence of people we first met in a religious setting? Do religious objects bring religious ideas to mind and therefore influence decisions made in the presence of religious material culture? In the search for individual-level effects of religion, situations may be better units of analysis than individuals.
Fourth, we might look harder for short-term causal effects of religion on individuals.
Experimental research consistently finds that bringing certain identities or norms or expectations to mind at the moment of decision-making changes behavior, but these behavioral changes often are situational and short-term. Dan Ariely’s research on cheating has shown that if people are reminded of the Ten Commandments immediately before they have an opportunity to cheat or steal, they are less likely to cheat or steal right at that moment (Ariely 2008:206-08). But they are no less likely to cheat or steal the next day, and being a generally religious person does not change that (Norenzayan and Shariff 2008). Lest we are tempted to commit the religious congruence fallacy in interpreting the Ten Commandments effect, the same reduction in cheating occurs if people engage in any sort of explicit moral reflection, religious or secular--or, for that matter, if they look in a mirror--immediately before being given an opportunity to cheat. The short-term, situational causal mechanisms found by social psychologists, cognitive scientists, and behavioral economists contrast sharply with the chronic, dispositional causal force that we often impute to religious ideas and practices.
Fifth, we might pay more attention to medium-term causal effects. We often do laboratory experiments that detect causal mechanisms operating in seconds and minutes, and we now do long-term panel studies looking for evidence of causal mechanisms operating over years.
But we do not often look for evidence of causal mechanisms that operate on the order of hours or days or weeks or months. There are research designs--for example, from the world of program
27 evaluation--aimed at finding causal effects on a time scale of days or weeks or months, but such designs are underused in the study of religion. We should give higher priority to investigating, say, whether attending church on Sunday morning makes you act any differently literally on
Sunday afternoon than to investigating whether generally attending church makes you generally act any differently. Religious incongruity implies that many common causal mechanisms involving religion probably operate in the short- and medium-term.
Sixth, when we find short- or medium-term causal effects we might ask how quickly they decay. Does participating in a religious ritual or worship service change your behavior one hour later? If so, is the effect still there two hours later? Six hours later? The next day? The next week? If people act differently when in the presence of a Bible or a WWJD bracelet or a copy of the Ten Commandments, how long does that difference last after they put down the Bible, take off the bracelet, or turn away from the wall displaying the Commandments? If you act differently when you first wear a WWJD bracelet, how long does it take before that effect wears off, and the bracelet becomes just another piece of mundane jewelry? When we find short- and medium-term effects, we should investigate their duration or half-life.
There are two related themes running through these suggestions. First, we should move towards a more deeply situational model of religious influence. We know well that religiosity is multidimensional, but we might better appreciate that most religion also is situational both in its everyday expression and in its everyday causal impact on individuals. Second, we should base our explanations and interpretations on accurate psychology.
28
Conclusion
My message is a simple one. We know that religious congruence is rare, so we should not expect to find it under every steeple, yet this knowledge does not inform our thinking and our practice as centrally as it should.
There is an irony here. I would like practice in the scientific study of religion to be more congruent with our knowledge, but I want us to more congruently apply our knowledge about how hard it is to be congruent!
There is an irony here, but not a contradiction. I want to leave you with the thought that the scientific study of religion can be an exception to the general rule that cultural systems are not congruent. As I have emphasized, congruence can be achieved, but only under special circumstances. The scientific study of religion, like any healthy scientific culture, can achieve more congruence than usually is possible in other domains because we work within institutions that promote congruence: peer review, publicly accessible data, forums for open debate, training programs in which students repeatedly have experiences that we hope create the right kind of internalized intellectual habits, and moments like this when we reflect together on our assumptions and on the relationship between our assumptions and our practices. The congruence that is realistically achievable within well functioning scientific communities is not usually achievable elsewhere because the social arrangements and institutions that promote congruence in science do not have easy functional equivalents in other domains.
So there is irony here, but no contradiction. The knowledge that only fools dance for rain in the dry season is a hard won discovery of the scientific study of religion. The same institutions that produced that discovery can help us remember it when we do our work.
29
REFERENCES
Adamczyk, A. 2009. Understanding the Effects of Personal and School Religiosity on the
Decision to Abort a Premarital Pregnancy. Journal of Health and Social Behavior
50(June):180-95.
Ariely, D. 2008. Predictably Irrational: The Hidden Forces that Shape Our Decisions. New
York: HarperCollins Publishers.
Batson, C. D. and A. Powell. 2003. Altruism and Prosocial Behavior. Pp. 463-84 in Handbook
of Psychology: Personality and Social Psychology, v. 5, edited by T. Millon and M. J.
Lerner. Hoboken, NJ: John Wiley & Sons.
Converse, P. E. 1964. The Nature of Belief Systems in Mass Publics. Pp. 206-61 in Ideology and
Discontent, edited by D. E. Apter. New York: Free Press.
D’Andrade, R. 1995. The Development of Cognitive Anthropology. New York: Cambridge
University Press.
Demerath, N. J. 2001. Crossing the Gods: World Religions and Worldly Politics. New
Brunswick, NJ: Rutgers University Press.
DiMaggio, P. 1997. Culture and Cognition. Annual Review of Sociology 23:263-87.
Dennett, D. C. 2006. Breaking the Spell: Religion as a Natural Phenomenon. New York: Viking.
Emerson, M. O. and C. Smith. 2000. Divided by Faith: Evangelical Religion and the Problem of
Race in America. New York: Oxford University Press.
Evans-Pritchard, E. E. 1965. Theories of Primitive Religion. Oxford: Clarendon Press.
Floyd, C. 2000. Daniel Dennett’s Darwinian Mind: An Interview with a ‘Dangerous’ Man.
Search (May/June). Retrieved from www.searchmagazine.org on 11/19/08.
30
Gallagher, S, K. and C. Smith. 1999. Symbolic Traditionalism and Pragmatic Egalitarianism:
Contemporary Evangelicals, Families, and Gender. Gender & Society 13(2):211-33.
Gallup, G. Jr., and D. M. Lindsay. 1999. Surveying the Religious Landscape: Trends in U.S.
Beliefs. Harrisburg, PA: Morehouse Publishing.
Gilkes, C. T. 1980. The Black Church as a Therapeutic Community: Suggested Areas for
Research into the Black Religious Experience. Journal of the Interdenominational
Theological Center 8(1):29-44.
Greeley, A. and M. Hout. 2006. The Truth about Conservative Christians: What They Think and
What They Believe. Chicago: University of Chicago Press.
Hitchens, C. 2007. God is Not Great: How Religion Poisons Everything. New York: Twelve.
Hitlin, S. 2008. Moral Selves, Evil Selves: The Social Psychology of Conscience. New York:
Palgrave Macmillan.
Keister, L. A. 2008. Conservative Protestants and Wealth: How Religion Perpetuates Asset
Poverty. American Journal of Sociology 113(5):1237-71.
King, M. 2009. Religious Belief and Practice: Good for You in the United States, Less Good in
Europe. Lecture delivered on 7 May 2009 at the Center for Spirituality, Theology, and
Health, Duke University. Video clip accessed on 10/9/09 at
www.spiritualityandhealth.duke.edu/education/sth-seminars.html.
Leach, M. M., M. E. Berman, and L. Eubanks. 2008. Religious Activities, Religious Orientation,
and Aggressive Behavior. Journal for the Scientific Study of Religion 47(2):311-19.
Lichterman, P. 2008. Religion and the Construction of Civic Identity. American Sociological
Review 73(1):83-104.
Lord, C. G. 1997. Social Psychology. New York: Harcourt Brace College Publishers.
31
Maio, G. R., J. M. Olson, M. M. Bernard, and M. A. Luke. 2003. Ideologies, Values, Attitudes,
and Behavior. Pp. 283-308 in Handbook of Social Psychology, edited by J. Delamater. New
York: Kluwer Academic/Plenum Publishers.
Moaddel, M. 2007. Introduction: Theoretical and Methodological Issues in the Study of Values.
Pp. 1-22 in Values and Perceptions of the Islamic and Middle Eastern Publics, edited by M.
Moaddel. New York: Palgrave Macmillan.
Myerhoff, B. 1978. Number Our Days. New York: Simon and Schuster.
Neitz, M. J. 2004. Gender and Culture: Challenges to the Sociology of Religion. Sociology of
Religion 65(4):391-402.
Nelson, T. J. 2005. Every Time I Feel the Spirit: Religious Experience and Ritual in an African
American Church. New York: New York University Press.
Norenzayan, A. and A. F. Shariff. 2008. The Origin and Evolution of Religious Prosociality.
Science 322 (3 October):58-62.
Olsson, T. 1999. Verbal Representation of Religious Beliefs: A Dilemma in the Phenomenology
of Religions. Pp. 75-92 in Comparative Studies in History of Religions: Their Aim, Scope,
and Validity, edited by E. R. Sand and J. P. Sørensen. Copenhagen: Museum Tusculanum
Press.
Palmer, C. T. and L. B. Steadman. 2004. With or Without Belief: A New Approach to the
Definition and Explanation of Religion. Evolution and Cognition 10(1):138-47.
Perrin, R. D. 2000. Religiosity and Honesty: Continuing the Search for the Consequential
Dimension. Review of Religious Research 41(4):534-44.
Poloma, M. P. 1989. The Assemblies of God at the Crossroads: Charisma and Institutional
Dilemmas. Knoxville: University of Tennessee Press.
32
Pope, L. 1942. Millhands & Preachers: A Study of Gastonia. New Haven: Yale University Press.
Regnerus, M. D. 2007. Forbidden Fruit: Sex & Religion in the Lives of American Teenagers.
New York: Oxford University Press.
Smilde, D. 2007. Reason to Believe: Cultural Agency in Latin American Evangelicalism.
Berkeley: University of California Press.
Stewart-Williams, S. and J. Podd. 2004. The Placebo Effect: Dissolving the Expectancy Versus
Conditioning Debate. Psychological Bulletin 130(2):324-40.
Swidler, A. 1986. Culture in Action: Symbols and Strategies. American Sociological Review
51(April):273-86.
Tambiah, S. J. 1990. Magic, Science, Religion, and the Scope of Rationality. New York:
Cambridge University Press.
Trimble, D. E. 1997. The Religious Orientation Scale: Review and Meta-analysis of Social
Desirability Effects. Educational and Psychological Measurement 57:970-86.
Tsitsos, W. 2007. The Interaction of Race and Theological Orientation in Congregation Social
Service Provision. Doctoral Dissertation. Tucson: Department of Sociology, University of
Arizona.
Vaisey, S. 2009. Motivation and Justification: A Dual-Process Model of Culture in Action.
American Journal of Sociology 114(6):1675-1715.
Wacker, G. 2001. Heaven Below: Early Pentecostals and American Culture. Cambridge:
Harvard University Press.
Wuthnow, R. 1997. The Cultural Turn: Stories, Logics, and the Quest for Identity in American
Religion. Pp. 245-65 in Contemporary American Religion: An Ethnographic Reader, edited
by P. E. Becker and N. L. Eiesland. Walnut Creek, CA: AltaMira Press.
33